MARINE STRATEGY FRAMEWORK
DIRECTIVE
Task Group 4 Report
Food webs
APRIL 2010
S. Rogers, M. Casini, P. Cury, M. Heath, X. Irigoien, H. Kuosa, M. Scheidat,
H. Skov, K. Stergiou, V. Trenkel, J. Wikner & O. Yunev
Joint Report
Prepared under the Administrative Arrangement between JRC and DG ENV
(no 31210 – 2009/2010), the Memorandum of Understanding between the
European Commission and ICES managed by DG MARE, and JRC’s own in-
stitutional funding
Editor: H. Piha
EUR 24343 EN - 2010
The mission of the JRC is to provide customer-driven scientific and technical support for the con-
ception, development, implementation and monitoring of EU policies. As a service of the Euro-
pean Commission, the JRC functions as a reference centre of science and technology for the Union.
Close to the policy-making process, it serves the common interest of the Member States, while
being independent of special interests, whether private or national.
European Commission
Joint Research Centre
Institute for Environment and Sustainability
Contact information
Address: Via Enrico Fermi 2749, 21027 Ispra (VA), Italy
E-mail: [email protected]
Tel.: +39 0332 785702
Fax: +39 0332 789352
International Council for the Exploration of the Sea
Conseil International pour l’Exploration de la Mer
General Secretary
H. C. Andersens Boulevard 44–46, DK-1553 Copenhagen V, Denmark
Tel.: +45 33 38 67 00
Fax: +45 33 93 42 15
www.ices.dk, [email protected]
Legal Notice
Neither the European Commission nor any person acting on behalf of the Commission is responsi-
ble for the use which might be made of this publication.
This report does not necessarily reflect the view of the European Commission and in no way an-
ticipates the Commission’s future policy in this area.
The views expressed in the report are those of the authors and do not necessarily represent the
views of ICES.
Europe Direct is a service to help you find answers to your questions about the European Union
Freephone number (*):
00 800 6 7 8 9 10 11
(*) Certain mobile telephone operators do not allow access to 00 800 numbers or these calls may be billed.
A great deal of additional information on the European Union is available on the Internet.
It can be accessed through the Europa server http://europa.eu/
JRC 58110
EUR 24343 EN
ISBN 978-92-79-15656-4
ISSN 1018-5593
DOI 10.2788/87659
Luxembourg: Office for Official Publications of the European Communities
© European Union and ICES, 2010
Reproduction is authorised provided the source is acknowledged
Printed in Italy
http://europa.eu.int/citizensrights/signpost/about/index_en.htm#note1#note1
PREFACE
The Marine Strategy Framework Directive (2008/56/EC) (MSFD) requires that the Euro-
pean Commission (by 15 July 2010) should lay down criteria and methodological stan-
dards to allow consistency in approach in evaluating the extent to which Good
Environmental Status (GES) is being achieved. ICES and JRC were contracted to provide
scientific support for the Commission in meeting this obligation.
A total of 10 reports have been prepared relating to the descriptors of GES listed in Annex
I of the Directive. Eight reports have been prepared by groups of independent experts co-
ordinated by JRC and ICES in response to this contract. In addition, reports for two de-
scriptors (Contaminants in fish and other seafood and Marine Litter) were written by
expert groups coordinated by DG SANCO and IFREMER respectively.
A Task Group was established for each of the qualitative Descriptors. Each Task Group
consisted of selected experts providing experience related to the four marine regions (the
Baltic Sea, the North-east Atlantic, the Mediterranean Sea and the Black Sea) and an ap-
propriate scope of relevant scientific expertise. Observers from the Regional Seas Conven-
tions were also invited to each Task Group to help ensure the inclusion of relevant work by
those Conventions. A Management Group consisting of the Chairs of the Task Groups
including those from DG SANCO and IFREMER and a Steering Group from JRC and
ICES joined by those in the JRC responsible for the technical/scientific work for the Task
Groups coordinated by JRC, coordinated the work. The conclusions in the reports of the
Task Groups and Management Group are not necessarily those of the coordinating organi-
sations.
Readers of this report are urged to also read the report of the above mentioned Manage-
ment Group since it provides the proper context for the individual Task Group reports as
well as a discussion of a number of important overarching issues.
Contents
Executive Summary ................................................................................................................. 1
1. Definition of terms, and scientific understanding of the key concepts associated
with Food Webs .......................................................................................................... 1
2. Good Environmental Status of food webs .................................................................. 1
3. How should “scale” be addressed ............................................................................... 2
4. Key Attributes of the Descriptor ................................................................................. 2
4.1. Attribute 1; Energy flows in food webs ...................................................................... 2
4.1.1. Description of attribute and why it is important 2
4.1.2. Indicators of the attribute 3
4.2. Attribute 2; Structure of food webs (size and abundance) .......................................... 3
4.2.1. Description of attribute and why it is important 3
4.2.2. Criteria: characteristics of the attribute with respect to GES 3
4.2.3. Indicators of the attribute 4
5. Method for aggregating indicators within the Descriptor to achieve an overall
assessment, if available ............................................................................................... 4
6. Emergent messages about monitoring and research, and Final Synthesis .................. 4
Report..................................................................................................................................... .. 5
1. Definition of terms ...................................................................................................... 5
1.1. Definition of key terms in descriptor .......................................................................... 5
1.2. Glossary of key terms in descriptor ............................................................................ 5
1.3. What is covered and what is outside scope ................................................................. 6
2. Scientific understanding.............................................................................................. 7
2.1. Good Environmental Status of Food Webs ................................................................ 7
2.2. Pressures acting on food webs .................................................................................... 7
2.3. Patterns in food webs .................................................................................................. 8
2.4. What is special about marine food webs ..................................................................... 8
2.5. Current considerations of food webs in management ................................................. 9
2.6. Existing approaches to monitoring Food Webs .......................................................... 9
2.7. Ecosystem Models ...................................................................................................... 9
3. Relevant spatial and temporal scales ........................................................................ 10
4. Key attributes of the descriptor ................................................................................. 10
4.1. Attribute 1: Energy flows in food webs .................................................................... 10
4.1.1. Criteria 1a) Production or biomass ratios that secure the long term viability of all
components 11
4.1.2. Criteria 1b) Predator performance reflects long-term viability of components 14
4.1.3. Criteria 1c) Trophic relationships that secure the long-term viability of
components 18
4.2. Attribute 2: Structure of food webs (size and abundance) ........................................ 21
4.2.1. Size based 21
4.2.2. Criteria 2a) Proportion of large fish maintained within an acceptable range 22
4.2.3. Abundance /distribution 23
4.2.4. Criteria 2b) Abundance /distribution maintained within an acceptable range 24
5. How are the indicators aggregated to assess GES for the descriptor? ...................... 25
5.1. Aggregation of assessments across Attributes .......................................................... 26
6. Emergent messages about monitoring and research and final Synthesis .................. 27
7. References ................................................................................................................. 28
8. Task group members ................................................................................................. 39
Annex 1. The role and merits of ecosystem models ............................................................... 41
Annex 2. Shellfish Production requirement ratio to support eider ducks (Somateria
mollissima) in the Wash, UK .................................................................................... 47
Annex 3. Changes in the food web of northwestern Atlantic shelf seas as a result of
collapse in cod stocks ............................................................................................... 48
Annex 4. Ratios of fishery landings (or derivatives of landings) to production of lower
trophic levels ............................................................................................................. 50
Annex 5. Background information on the Marine Trophic Index (MTI) ............................... 51
Annex 6. Abundance of species ............................................................................................. 54
| 1
EXECUTIVE SUMMARY
The 2008 European Marine Strategy Framework Directive (2008/56/EC) includes a re-
quirement for EU Member States to report on the environmental status of the seas under
their jurisdiction and to work to achieve Good Environmental Status (GES). This is de-
fined by eleven qualitative descriptors, and one of them deals with „Food Webs‟.
The Task Group 4 „Food Webs‟ descriptor reads: All elements of the marine food webs, to
the extent that they are known, occur at normal abundance and diversity and levels capa-
ble of ensuring the long-term abundance of the species and the retention of their full re-
productive capacity.
This report defines the terms used in this descriptor (section 2), describes the scientific
understanding (section 3) and the relevant spatial and temporal scales (section 4). A
framework to describe attributes of GES for food webs is provided in section 5.
1. DEFINITION OF TERMS, AND SCIENTIFIC UNDERSTANDING OF THE KEY CONCEPTS ASSOCIATED WITH FOOD WEBS
Food webs are networks of feeding interactions between consumers and their food. The
species composition of food webs varies according to habitat and region, but the principles
of energy transfer from sunlight and plants through successive trophic levels are the same.
This descriptor addresses the functional aspects of marine food webs, especially the rates
of energy transfer within the system and levels of productivity in key components.
‘All elements.’ All components of food webs have been considered, i.e. all trophic and
functional groups, comprising either one or several species. This potentially includes all
living organisms and non-living organic components.
‘..to the extent that they are known..‟ While examination of food webs should in principle
include „all elements‟, for practical purposes it would include only those food web compo-
nents that can effectively be sampled by established robust methods of monitoring.
„..normal abundance and diversity and at levels capable of ensuring the long-term abun-
dance of the species and the retention of their full reproductive capacity.‟ This provides
guidance on the reference points and/or target values selected to correspond to good envi-
ronmental status. Full reproductive capacity refers to the maintenance of fertility and
avoidance of reduction in population genetic diversity.
2. GOOD ENVIRONMENTAL STATUS OF FOOD WEBS
The interactions between species in a food web are complex and constantly changing,
making it difficult to identify one condition that represents „good‟ status. However,
changes in species relative abundance in an ecosystem will affect interactions in several
parts of a food web, and may have an adverse effect on food web status. There is, how-
ever, a significant lack of understanding to assess the ecosystem consequences of such
change, or the value that society should attribute to it. As all marine food webs have al-
ready been adversely affected by humans, a judgement will need to be reached by Member
States to identify regional limit reference points.
Good Environmental Status of Food Webs will therefore be achieved when the indicators
describing the various attributes of the descriptor reach the thresholds set for them. These
should ensure that populations of selected food web components occur at levels that are
within acceptable ranges that will secure their long-term viability. Components must be
selected carefully to avoid use of large numbers of species for which abundance / biomass
| 2
trends are required (i.e. avoid use of general terms such as „predators‟ or „prey‟). Assess-
ment of food webs will need to include;
(i) biological groups with fast turnover rates (e.g. phytoplankton, zooplankton, bacteria) that will respond quickly to system change;
(ii) groups that are targeted by fisheries;
(iii) habitat-defining groups; and
(iv) charismatic or sensitive groups often found at the top of the food web.
3. HOW SHOULD “SCALE” BE ADDRESSED
Attributes of food webs can in principle be applied on any spatial scale or time scale, how-
ever, there are clear interpretational and practical limitations. The fundamental time scale
over which ecosystem assessments might be required is annual. The temporal scale neces-
sary to assess growth, mortality and feeding fluxes between food web components should
be annual to integrate over seasonal variability at the lowest trophic levels. More frequent
assessments, for example those that could be undertaken monthly, are operationally diffi-
cult to undertake and maintain, and their interpretation becomes complicated by seasonal
dynamics. For the higher trophic levels, some smoothing of annual rates may be required
to eliminate inter-annual variability. For longer lived species such as piscivorous fish,
mammals and birds, assessments on an annual basis may be too frequent since variability
at this scale becomes more influenced by unexplained processes such as recruitment vari-
ability, and less by internal population processes.
Similar issues apply to considerations of appropriate spatial scales: at small spatial scales,
such as parts of a MSFD Sub-Region, immigration and emigration by advection and mi-
grations become important components of change. For large, long-lived taxa, spatial scales
which integrate over migration ranges may be appropriate, but these scales may span fun-
damentally different habitats and communities for lower trophic levels, for example plank-
ton or benthos, to the point that a synthesis at this scale becomes questionable.
4. KEY ATTRIBUTES OF THE DESCRIPTOR
The effects of fishing are the most important pressures which directly affect target species,
and indirectly affect other non-target components of food webs. While these effects re-
spond to management action, the components which they influence are also subject to cli-
mate variation and other natural drivers making precise attribution of cause and effect
difficult. It is also likely that other pressures will need to be considered in the development
of measures, and particularly the cumulative effects of multiple activities.
4.1. Attribute 1; Energy flows in food webs
4.1.1. Description of attribute and why it is important
The food web is a fully interconnected system, so pressures on one part of the system may
have impacts elsewhere which are not easily predictable. For example, harvesting of san-
deels in the North Sea, where they are a key species in the food web, will remove food for
birds, mammals, piscivorous fish, and release predation pressure on zooplankton. There
may also be indirect consequences for a range of other species. Managing human activity
to achieve a desired balance between species in the system is therefore a major challenge.
Energy flows through the food web are an attribute which allows us to diagnose the state
of the system.
| 3
4.1.2. Indicators of the attribute
We identify three criteria of energy flows in the food web which are feasible to measure
and apply at a regional scale: a) ratios of production at different trophic levels, b) the pro-
ductivity (production per unit biomass) of key species or groups, and c) trophic relation-
ships. Many indicators within each criterion require further elaboration to become
operational, and it is not yet possible to robustly define thresholds or limit reference points,
or the full extent to which climate change may affect the metrics.
a) Production or biomass ratios that secure the long term viability of all components. Ra-
tios of production or biomass between different trophic levels in the food web provide
measures of the pattern of energy flow, and the efficiency of energy transfer through the
web. It is proposed that a ratio indicator is developed, specific to each marine Regions or
Sub-Regions, and based on either ratios of pelagic to demersal fish biomass and/or produc-
tion, or benthos to fish production, or the proportions of plankton and benthos production
required to support fisheries.
b) Predator performance reflects long-term viability of components. Some species, or
groups of species, may act as guides to change in the ecosystem. The performance of these
species, as measured by their productivity, effectively summarises the main predator-prey
processes in the neighbourhood of the food web that they inhabit. The basis for such
measures is already established in OSPAR EcoQO, for example in terms of the fledging
success of kittiwakes, which relates to the availability of sandeels. Following the same
principle, we propose indicators based on the nutritional status of marine mammals or sea-
birds.
c) Trophic relationships that secure the long-term viability of components.
The diet composition of a group of species is dependent on the consumption by each com-
ponent species and can be a valuable measure of the relative abundance of prey in a food
web and the degree of connectivity in the food web. The diet of some single species, par-
ticularly top predators, can provide similar insights. For group-level assessment, the Ma-
rine Trophic Index has been used to calculate the mean feeding level of a group from
species composition data, assuming a particular diet for each species. At the species level,
changes in stomach contents (which indicate the trophic level of diet) can also be diagnos-
tic of underlying change in the food web.
4.2. Attribute 2; Structure of food webs (size and abundance)
4.2.1. Description of attribute and why it is important
Size structure of food webs is an important attribute and integral to the maintenance of
predator prey relationships. Most life history traits are correlated with size, which con-
strains metabolic rate and controls growth, reproduction and survival, so body size is also a
proxy for trophic level. Fishing is usually size-selective within species, so larger individu-
als generally suffer greater rates of mortality. Exploited populations and communities
therefore contain relatively fewer large fish and mean size is reduced. This may in turn
have an indirect impact on their prey populations as a result of size-dependent predation
and changes in density-dependent growth. The abundance (and distribution) of carefully
selected indicator populations (e.g. jellyfish, plankton, etc) can describe food web status
and/or levels of human perturbation.
4.2.2. Criteria: characteristics of the attribute with respect to GES
Changes in the mean size of fish and the proportion of large species in the community can
be detected by indicators of the mean size and size distribution. It is, however, difficult to
determine reference values for size-based community indicators. Attempts to do so have
| 4
been based on modelling the expected community structure in the absence of fishing, or by
selecting a time in the past when the community structure was judged to have been accept-
able.
Changes in absolute or relative abundance can be assessed in relation to reference direc-
tions and limit reference points, rather than specific targets. For many species, minimum
viable populations can be inferred from ecosystem models.
4.2.3. Indicators of the attribute
Monitoring the rate of change of functionally important species to highlight rapid in-
creased or decreased abundance will help to identify where future management action may
be required. The following two criteria are proposed;
a) Proportion of large fish maintained within an acceptable range. This criterion describes
the changes in the proportion of large fish, and hence the average weight and average
maximum length of the fish community in a Region or Sub-Region. The OSPAR EcoQO
(Proportion of large fish), provides a protocol that can be applied in other regional seas.
b) Abundance maintained within an acceptable range; To make this criterion operational
requires an assessment of the most suitable species in a Region or Sub-Region to represent
food web integrity, based on key biological groups present. Indicators should describe re-
gional abundance trends to identify changes in population status that may have implica-
tions for food web status.
5. METHOD FOR AGGREGATING INDICATORS WITHIN THE DESCRIPTOR TO ACHIEVE AN OVERALL ASSESSMENT, IF AVAILABLE
TG4 identifies two main attributes of food webs, „Energy flows in food webs‟ and „Struc-
ture of food webs (size and abundance)‟. It is necessary that both attributes must be ad-
dressed for an assessment to be acceptable. Within each attribute TG4 suggests a number
of promising criteria, but there may be others. To overcome the burden of proof within an
attribute, it will be necessary to address the entire spatial extent of the assessment Region
or Sub-Region. This can be achieved using a suite of localised indicators which together
cover the domain, or a single spatially comprehensive indicator. More work is required to
understand the practical implications of this requirement for Member States or Regional
Seas Conventions.
6. EMERGENT MESSAGES ABOUT MONITORING AND RESEARCH, AND FINAL SYNTHESIS
There are several operational indicators already in use that are relevant to this descriptor of
GES, and that can contribute to the assessment of food web dynamics. It is encouraging to
note that these are coherent with other international activities to ensure sustainable fisher-
ies and maritime strategy in European waters, therefore allowing coordinated activity by
Member States. While it is therefore possible to begin work now, some further develop-
ment is required for indicators that cover all the criteria identified in TG4.
The practical process for achieving GES for this descriptor is not well defined. The com-
pletion of monitoring programmes and delivery of food web indicators for a Regional Sea
in which several Member States have a stake will require substantial levels of coordina-
tion. This will have a major influence on successful implementation of the Directive.
| 5
1. DEFINITION OF TERMS
1.1. Definition of key terms in descriptor
‘All elements of the marine food webs,..’
The structure of food webs is generically the same as they all involve predator prey inter-
actions and energy transfer between levels, but the species composition of food webs var-
ies according to the environment in which they occur. Food webs in different regions are
therefore distinguished by interactions between key species, but the processes of energy
transfer are the same. This description defines the spatial scale of food webs used in this
report.
We interpret all „elements‟ as all food web components, i.e. all trophic and functional
groups, which could be made up of one or several species. This includes living organisms
(from higher predators such as birds and marine mammals to bacteria and viruses) and
non-living components (detritus and dissolved nutrients).
„..to the extent that they are known..‟
This includes all food web components that can be sampled by established methods of
monitoring.
‘..occur at normal abundance and diversity and at levels capable of ensuring the long-
term abundance of the species and the retention of their full reproductive capacity.‟
Normal abundance should be interpreted as the reference point / target values selected to
correspond to good status. In the MSFD this represents a sustainable state of use from an
ecosystem perspective. For living organisms this is an abundance that can recover from
perturbation caused by human induced pressures within a reasonable time frame. A „Nor-
mal‟ assemblage is also interpreted as having a functional diversity that would be typical
for the marine region and under the prevailling conditions of climate to ensure the overall
functioning of the ecosystem.
Full reproductive capacity is not interpreted in the way that is defined in ecology (which is
the maximum lifetime reproductive output of a species). Full reproductive capacity refers
to the maintenance of fertility and avoidance of reduction in population genetic diversity.
Full reproductive capacity sustains the functions of the species in the assemblage.
1.2. Glossary of key terms in descriptor
A food web is a “Representation of feeding relationships in a community that includes all
the links revealed by dietary analysis” (Begon et al. 1995) (Figure 1-1). In other words it
describes those organisms that are eaten by other organisms. Parasitism and disease is in
principal a predator-prey interaction, but by smaller organisms (e.g. bacteria or virus) on
larger (e.g. phytoplankton or fish).
| 6
Figure 1-1 Simplified model of a marine food web. Size ranges of organisms indicated by the
numbers. Major trophic interactions are shown by arrows, as well as sedimentation of particulate
matter and excretion of nutrients. Groups with auto-, hetero- and mixotrophic organisms are
shown by green and white boxes. Cili.=ciliates, Flag.=flagellates, Fil. Cya.=filamentous cyanobacte-
ria.
Trophic group; refers to a category of organisms within a trophic structure, defined ac-
cording to their mode of feeding.
Functional Group; a group of organisms that are using the same type of prey.
1.3. What is covered and what is outside scope
This report will deal with only the functional aspects of marine food webs, especially the
rates and directions of energy transfer within the system and levels of productivity in key
components. This descriptor will generally not address structural indicators of biodiversity
for common benthic or pelagic communities using, for example, metrics of species relative
abundance or biomass. It is intended that the Biodiversity Task Group (TG 1) will deal
with these along with other measures of diversity relating to threatened, declining and
charismatic species. However the abundance and distribution of some key species or func-
tional groups (top predators, jellyfish etc) can be representative of substantial parts of food
webs, so where they are considered sufficiently important they will be included in this
descriptor. Abundance of non-indigenous species will be dealt with by TG 2, commercial
fish and shellfish stock status will be dealt with by TG 3, and the description of benthic
communities and biotic structure, substrate and habitat structure will be dealt with by the
group ad-dressing sea-floor integrity (TG 6).
| 7
2. SCIENTIFIC UNDERSTANDING
2.1. Good Environmental Status of Food Webs
This descriptor is one of three which addresses marine biodiversity. It is also one of the
most difficult to implement. The food webs descriptor deals with the functional aspects of
species interactions, especially the rates and directions of energy transfer within the system
and levels of productivity in key components. Metrics to describe food web status should
consider both the extent of bottom-up controls on marine ecosystems, as well as highlight-
ing top-down controls.
Such assessments will take account of the pressure exerted by top predators on prey com-
munities, using, for example, estimates of productivity, reproductive success and size-
based measures of population change. There has been recent progress to develop indicators
for some of these processes, including the development of OSPAR Ecological Quality
Objectives for seabirds and fish communities in the North Sea, and other applications
elsewhere in Europe. The productivity of primary and secondary producers has not been
included elsewhere in the GES descriptors yet is important to describe the functioning of
marine ecosystems. Thus the extent of plankton productivity using both field and remote
observation will be used to generate metrics that describe the food supply available for
dependent predators. The contribution of ecosystem and food web modelling will provide
useful insights into future scenarios of ecosystem change.
Attributing the cause of change in food web structure or function is complex, and will be
the result of pressures which act both directly and indirectly on different components of
the ecosystem. It will therefore be necessary, wherever possible, to develop metrics that
respond to a manageable activity, so that the assessment of good environmental status can
lead to specific monitoring requirements and appropriate thresholds or reference levels.
One of the most valuable contributions that can be made by the descriptor „Food webs‟ is
to provide an overview of broad scale ecosystem status, integrating across a number of
different trophic groups, and usually at a broad scale. This is a distinguishing feature of the
descriptor, and compliments those also focussing on biodiversity issues (descriptors 1, 2
and 6). The spatial scale at which food web status is monitored is likely to reflect local or
regional environmental conditions, and be dependent on the availability of data for key
components.
The interactions between species in a food web are complex and constantly changing,
making it difficult to identify one condition that represents „good‟ status. However,
changes in species relative abundance in an ecosystem will affect interactions in several
parts of a food web, and may have an adverse effect on food web status. There is, how-
ever, a significant lack of understanding to assess the ecosystem consequences of such
change, or the value that society should attribute to it. As all marine food webs have al-
ready been adversely affected by humans, a judgement will need to be reached by Member
States to identify regional limit reference points.
Good Environmental Status of Food Webs will therefore be achieved when the indicators
describing the various attributes of the descriptor reach the thresholds set for them. These
should ensure that populations of selected food web components occur at levels that are
within acceptable ranges that will secure their long-term viability.
2.2. Pressures acting on food webs
Patterns in the structure and function of marine ecosystems can be substantially affected
by both environmental changes (e.g. through interannual and interdecadal climatic varia-
tion and change) and the pressures of human activities such as fishing effects (e.g. through
| 8
overexploitation of large predatory or forage fishes) (Cury et al., 2003). Different types of
controls can therefore be exerted on marine ecosystems and can lead to alternate states.
The effects of fishing are the most important pressures which directly affect target species,
and indirectly affect other non-target components of food webs. While these effects re-
spond directly to management action, the response time can be slow and variable, and re-
covery can be impeded by the influence of other natural drivers, making precise attribution
of cause and effect difficult.
2.3. Patterns in food webs
Bottom-up control is the conventional trophic flow control that seems to dominate most
ecosystems, where the regulation of food-web components derives from change in the
abundance of primary producers which is itself strongly influenced by environmental con-
ditions. Literature documenting the relationship between the abundance of different tro-
phic levels and environmental variability is widely available. This has been documented
for example in the North Atlantic where parallel long-term trends across four marine tro-
phic levels, ranging from phytoplankton, zooplankton and herring to marine birds, have
been related to environmental changes in the North Sea (Aebisher et al., 1990).
Top-down control is the regulation in abundance that is exerted by predators on their prey.
A large reduction in predator abundance can cause an increase in prey that cascades
downward in the food chain, a phenomenon known as a trophic cascade. Trophic cascades
can therefore be thought of as reciprocal predator–prey effects that alter the abundance,
biomass or productivity of a population or trophic level across more than one link in a food
web, resulting in alternate trends between different trophic levels. The decline in top
predator abundance has been demonstrated to cascade down several marine food webs.
Recent studies reveal that reduced abundance of large fish predators (e.g. cod) had pro-
found effects on the abundance of small pelagic fishes which in turn affect plankton dy-
namics in the Black Sea, the NW Atlantic ecosystems and the Baltic sea (Casini et al.,
2008). Substantial reductions in marine mammal, shark, and piscivorous fish abundance
have led to increased abundances of mesopredators and invertebrate predators. Predation
has also inhibited recovery of depleted species, sometimes through predator–prey role re-
versals.
Top-down and bottom-up processes are not mutually exclusive within ecosystems. In fact,
both ways of ecosystem control may act in concert and their relative strength can vary in
response to ecosystem alterations (Litzow and Ciannelli, 2007; Casini et al., 2009).
In several productive upwelling ecosystems (e.g. Canary, Benguela, California and Hum-
boldt currents), there is an intermediate trophic level, occupied by a limited number of
species of small, plankton-feeding pelagic fish, comprising substantial populations that are
exploited intensively and vary considerably in abundance (Cury et al., 2000). Examples
are capelin in the Norwegian Sea, anchovy or sardine in some upwelling systems. Pelagic
fish can exert a major control on energy flows in productive ecosystems, and this has been
termed „wasp-waist‟ control as those forage fish resources can affect trophic levels both
downwards and upwards (i.e. a bottom-up control of top predators by small pelagic fishes,
and top-down control of plankton by Small pelagic fishes). The collapse of small pelagic
fish populations in the northern Benguela had profound effects on top predators such as
marine bird and mammals as well as on lower trophic levels such as jelly-fish (Cury and
Shannon, 2004).
2.4. What is special about marine food webs
The structure of marine food webs is not inherently different from terrestrial or freshwater
ecosystems, so classical food web theory also applies to marine systems. This theory is
| 9
most relevant for conservation biology, specifically related to biodiversity issues (May,
2009) rather than the management of exploited populations. Marine food webs are, how-
ever, characterised by many weak links between species and relatively short average path
lengths (Link, 2002). This high level of connectance in most marine food webs makes
them relatively robust to the secondary effects of species declines or local extinction. Short
average path lengths between species suggest that perturbations such as fishing or climate
change will be transmitted more widely throughout marine ecosystems compared to their
terrestrial or freshwater counterparts (Dunne et al., 2004). Furthermore, body size is an
important structuring variable in marine communities and consequently size spectra have
been much studied in marine systems, though the implications for food web functioning
have been less well studied (Jennings et al., 2001; Raffaelli et al., 2005).
2.5. Current considerations of food webs in management
Food web issues are of increasing importance in European marine management and legis-
lation, though there are few tools or frameworks in current use that focus on food webs or
relationships between species. The main approaches use multi-species models for deter-
mining maximum sustainable yield values (multi-species MSY). These developments have
been driven by criticism of the long-standing single-species approach to European fishe-
ries management which take no account of the state of prey and predator populations.
Worm et al. (2009) analyzed current trends in multi-species exploitation rates and biomass
in a range of well studies fisheries ecosystems using Ecopath/Ecosim and „Atlantis‟ mod-
els. In 5 of 10 well-studied ecosystems, the average exploitation rate has recently declined
and is now at or below the rate predicted to achieve maximum sustainable yield for
seven
systems. Yet 63% of assessed fish stocks worldwide still require rebuilding, and even
lower exploitation rates are needed to reverse the collapse of vulnerable species. Crucially,
the sum of single-species MSY was generally a poor predictor of multi-species MSY. This
is thought to be because of difficulty in deciding a priori whether depensatory or compen-
satory responses to fishing will occur as a result of food web interactions. Each response
will lead to divergence between yields at the system level and those predicted by single-
species assessments. Similar results were found for the Eastern Bering Sea/Gulf of Alaska
(Mueter and Megrey, 2006), and the North Sea (Mackinson et al., 2009). Overall, the sum
of predicted single-species MSY differed from system-level MSY by more than 20% in
42% of the systems and by more than 50% in 18% of the systems analysed by Worm et al.
(2009).
2.6. Existing approaches to monitoring Food Webs
The extent to which communities function normally depends on the trophic structure and
size structure of their component taxa. A number of metrics have been proposed for moni-
toring these functions in marine communities (e.g. Rochet and Trenkel, 2003; Cury et al.,
2005). The metrics fall into several categories: a) assessment of the biomass/abundance of
trophic groups or ratios of groups (total/mean weight/abundance or mean trophic level of
piscivores, planktivores, benthivores), b) metrics derived from size structure (slope of size
spectrum, mean length), and c) metric describing linkages or networks (consumption ratio,
number of cycles in food web, mean number of trophic links). No reference points with a
theoretical basis exist currently for any of these metrics, though Link (2005) proposed
limit and warning values for some of them, and other authors have suggested the use of
reference directions (Jennings and Dulvy, 2005, Rochet et al., 2005).
2.7. Ecosystem Models
The multitude of links and processes that make up a real food web mean that the conse-
quences of change will probably be much wider than expected and, because of non-linear
| 10
relationships between species, may even lead to counterintuitive outcomes. Trophic eco-
system models are an important component of the tools that will be needed to advise on
the state of food webs and the extent of impacts. However, these models are still in an ear-
ly stage of development and the strengths and weaknesses of the various alternatives are
difficult to understand, though there have been some important reviews (Fulton et al.,
2005). In general, there seems to be a humped relationship between the detail included in a
model, and its effectiveness. Too little detail is ineffective because the model is too ab-
stract. Too much detail is ineffective because the model tries to capture all known
processes but at the expense of requiring detail for too many poorly understood parame-
ters. Between these extremes is a set of models that can be parameterised and will effec-
tively at represent the key properties of the system. An important recommendation from
reviews is that the use of a single ecosystem model is ill-advised. The comparative and
confirmatory use of multiple „minimum-realistic‟ models is strongly recommended. More
detailed information in the role and merits of ecosystem models is given in Annex 1.
3. RELEVANT SPATIAL AND TEMPORAL SCALES
Attributes of food webs can in principle be applied on any spatial scale or time scale, how-
ever, there are clear interpretational and practical limitations. The fundamental time scale
over which ecosystem assessments might be required is annual. The temporal scale neces-
sary to assess growth, mortality and feeding fluxes between food web components should
be annual to integrate over seasonal variability at the lowest trophic levels. More frequent
assessments, for example those that could be undertaken monthly, are operationally diffi-
cult to undertake and maintain, and their interpretation becomes complicated by seasonal
dynamics. For the higher trophic levels, some smoothing of annual rates may be required
to eliminate inter-annual variability. For longer lived species such as piscivorous fish,
mammals and birds, assessments on an annual basis may be too frequent since variability
at this scale becomes more influenced by unexplained processes such as recruitment vari-
ability, and less by internal population processes.
Similar issues apply to considerations of appropriate spatial scales: at small spatial scales,
such as parts of a MSFD Sub-Region, immigration and emigration by advection and mi-
grations become important components of change. For large, long-lived taxa, spatial scales
which integrate over migration ranges may be appropriate, but these scales may span fun-
damentally different habitats and communities for lower trophic levels, for example plank-
ton or benthos, to the point that a synthesis at this scale becomes questionable. Ultimately,
it seems likely that the appropriate spatial scale at which to assess food webs will be set by
the purpose for which the assessment is required rather than any ecological considerations.
Other practical considerations, such as the availability and spatial extent of monitoring
data for key taxa, are also likely to influence the scale at which assessments are made.
4. KEY ATTRIBUTES OF THE DESCRIPTOR
Based on current understanding of food web trophodynamics and the key components that
are available for study, it was agreed that fundamental attributes of food webs related to
the flow of energy (as carbon) through the system, and the structural features of compo-
nents, specifically their size and abundance. The following section introduces each of these
attributes and suggests criteria that might be applied to determine their status.
4.1. Attribute 1: Energy flows in food webs
Meta-analyses of marine ecosystems show a generic relationship between primary produc-
tion (standardised to sea surface area), and production at successively higher trophic lev-
| 11
els, for example fish (Nixon 1988; Iversen, 1990; Chassot et al., 2007). This relationship is
an expression of the efficiency with which the energy captured by primary production is
transferred up the food web. Within individual regions this efficiency may change over
time depending on a variety of human interventions and climatic factors. The aim of the
approach described in this section is to summarise the energy flow by means of a set of
metrics which allow an assessment of efficiency, and ultimately to allow an assessment of
whether there is unacceptable damage by human activity.
At some levels in a food web, energy flow may pass through a large number of predator-
prey linkages, whilst at others the flow may be focused through only a small number of
species and/or developmental stages or „bottlenecks‟ in the web. In some cases, bottle-
necks in the web may lead to a so-called „wasp-waist‟ food-web - one in which a single
species acts as a conduit between the lower and upper trophic levels. Such systems are
especially sensitive to changes in mortality of the key bottleneck species (Cury 2000),
which are thus key components for monitoring the state of the food web.
Metrics which aim to summarise energy flow through the system must incorporate, implic-
itly or explicitly, data from a number of different trophic levels. Absolute levels of primary
production, plankton, fish or seabird production, cannot in themselves be diagnostic of
flows. However, we can identify three generic types of measures that can be diagnostic of
energy flows and patterns: i) ratios of production at different trophic levels, ii) the produc-
tivity (production per unit biomass) of key species or groups, and iii) the trophic level of
the species or group of species.
i ) The concept of ratios of production is straightforward. The ratio of, for exam-ple, benthic to planktonic secondary production is a clear statement of the pro-
portion of primary production which is diverted to the benthic seabed food web
as opposed to the planktonic water column food web.
ii ) The diet of individual species in the food web will be largely determined by the abundances of suitable prey taxa to which they have access. Some predator
species, or groups of species, may play a significant part in food web dynamics
and thereby act as indicators of change in the system as a whole. The perform-
ance of these species, as measured by their productivity, effectively summarises
the main predator-prey processes in the neighbourhood of the food web that
they inhabit.
iii ) The diet composition of a group of species is dependent on the consumption by each component species and can be a valuable measure of the relative abun-
dance of prey in a food web and the degree of connectivity in the food web.
The diet of some single species, particularly top predators, can provide similar
insights. For group-level assessment, the Marine Trophic Index has been used
to calculate the mean feeding level of a group from species composition data,
assuming a particular diet for each species. At the species level, changes in
stomach contents or isotopic compositions (which indicate the trophic level of
diet) can also be diagnostic of underlying change in the food web.
4.1.1. Criteria 1a) Production or biomass ratios that secure the long term viability of all components
The purpose of applying ratios of production or biomass for assessing GES is to detect
gross structural changes in the energy flow through a food web which may have been
caused by, for example, removal of key species by harvesting, or disruption of distribu-
tional overlap between predators and prey through climatic factors. Examples of the type
of change which ratios of production would be intended to detect are: a) dominance of
jellyfish as planktivores in a system as a result of over-harvesting of small pelagic fish or
| 12
removal of top predators, or b) increased abundance of benthic invertebrates due to over-
harvesting of benthivorous demersal fish.
Fishing is an important pressure in most if not all shelf ecosystems, but not necessarily the
only factor which may shift energy flow between major pathways through the food web.
Hence it is important to remember that production ratios are snapshots summarizing mul-
tiple cumulative effects on the system. For example, trends in the ratio of macrobenthos to
demersal fish production in the North Sea imply top-down control of the benthos by fish
predation (Heath, 2005a) and explain the emergence of Nephrops fisheries as cod and oth-
er gadoid species have been depleted by harvesting. However, both environmental and
fishery changes have been suggested to be responsible for recently observed shifts in ben-
thic invertebrate to groundfish dominance and the emergence of shrimp fisheries in the
Gulf of Alaska (Anderson and Piatt, 1999; Bailey, 2000).
Many investigators have examined ratios between fishery yield/landings from an ecosys-
tem and the underlying primary production. Comparing across ecosystems, some consis-
tency in this ratio certainly exists (Nixon 1988; Iverson, 1990; Chassot et al., 2007;
Gaichas et al., 2009). However, ratios of bulk fishery yield to primary production take no
account of the species or functional group composition of catches. This is clearly impor-
tant since harvesting of high trophic level piscivore species accounts for more primary
production than harvesting low trophic level planktivores. For this reason, a more logical
approach is to compare the Annual Production Requirement of fishery catches resolved to
a given trophic level, with total production at that level. This ratio expresses the proportion
of production removed by fisheries. Annual Production Requirement is equivalent to the
term Primary Production Requirement (PPR) as defined by Pauly and Christensen (1995),
but not necessarily resolved to level of phytoplankton. The concept of an Annual Produc-
tion Requirement ratio can be applied in other ways than to fishery catches. For example,
the prey production required to support a given population of a predator can be derived
from an energetic model of that predator, and compared to the measured production of
prey in the environment. In this case the ratio expresses the contribution of the predator to
total utilisation of the prey. An example of this application of Production requirement ra-
tios is given in Annex 2.
Production or biomass ratios have been used effectively to identify fundamental characte-
ristic differences between ecosystems (Gaichas et al., 2009; Pranov and Link, 2009), how-
ever, setting management thresholds or limits to such ratios within a system is difficult.
The pelagic to demersal fish biomass ratio was considered to be one of the most robust
ecosystem indicators of fishing effects by Fulton et al. (2005). Based on experience in the
Georges Bank and in other heavily exploited systems, Link (2005) suggests that a warning
threshold has been crossed when pelagic fish biomass exceeds 75% or drops below 25% of
total fish biomass. However, other results are somewhat counter intuitive with respect to
the effect of fishing. For example, the pelagic to total fish biomass ratio in the Norwegian
Sea was found to be 0.85, despite the fact that fishery catch is low in this system relative to
others in the North Atlantic (Gaichas et al., 2009).
Assessments of fishery yield to primary production are typically undertaken on higher pre-
dators for which sampling is relatively simple (Nixon 1988; Gaichas et al., 2009). A simi-
lar promising indicator for food web efficiency at the base of the food web measures the
relative flow of biomass in the food web through the microbial heterotrophic component,
(Turley et al. 2000). This indicator has relevance for fish yield, sediment flux and thereby
also benthic production. It is therefore proposed for further development and evaluation,
and to be considered in future recommendations. Methods can be used in routine monitor-
ing programs at reasonable cost and with good spatio-temporal coverage. The measure is
based on bacterial community biomass production (e.g. 3H-thymidine uptake) relative to
| 13
autotrophic planktonic primary production (e.g. 14
HCO3- uptake method) (Turley et al.
2000).
4.1.1.1. Recommended production or biomass ratios
It is recommended one region-specific indicator is developed based on one of the follow-
ing examples;
Ratio of pelagic to demersal fish production or biomass. Annex 3 provides an example of
the application of pelagic to demersal fish biomass ratios to diagnose changes in the food
web of northwestern Atlantic ecosystems following the collapse of cod stocks. The ratio
has been identified as a robust indicator of food web status. Preliminary thresholds have
been suggested.
Ratio of macrobenthos invertebrate to demersal fish production or biomass. Annex 3 also
provides an example of the application of benthic invertebrate to demersal fish biomass
ratios to diagnose changes in the food web of northwestern Atlantic ecosystems following
the collapse of cod stocks. The ratio has also been identified as an indicator of major food
web impacts due to harvesting of demersal fish in the North Sea. No indicative thresholds
have been identified.
Ratio of zooplankton production requirement of landings to zooplankton production as a
measure of pressure on the food web due to fishing. An application of this ratio in the
North Sea, Celtic Sea and west of Scotland is given in Annex 4. The ratio indicates in-
creasing and more intense fishing pressure in the North Sea and west of Scotland area
compared to the Celtic Sea. Fishing pressure in the Celtic Sea appears lower than the other
area due to the lack of large scale industrial fisheries for small pelagic species.
Ratio of benthos requirements of landings to benthos production as a measure of pressure
on the food web due to fishing. An application of this ratio in the North Sea, Celtic Sea
and west of Scotland is given in Annex 4. The ratio indicates more extreme fishing pres-
sure in the west of Scotland area compared to the others. Pressure in the Celtic Sea appears
to be increasing due to escalating removals of high trophic level species and Horse Mack-
erel. The latter has a benthic component of diet and has increased in abundance in the re-
gion due to a move towards the poles in its‟ geographic range.
4.1.1.2. Technical evaluation of production of biomass ratio indicators
Ease of understanding
The overall concept is easy to understand and communicate.
Data availability
Much of the data required to derive ratios of abundance or production across a wide range
of trophic levels are already collected from fish assessment surveys, fishery landings,
plankton assessment surveys using e.g. the Continuous Plankton Recorder, and potentially
also remote sensing programmes.
The task of assembling data sets spanning a range of trophic levels or groups of species
would be a departure from the current working practices of scientific assessments for most
EU waters. For example, all ICES fish stock assessments are carried out on a species-by-
species basis, with no overview of the total pelagic or demersal fishery or survey data, or
of data from other trophic levels. Derivation of food web production or biomass ratios will
require the science community to take a wider view of data gathering and synthesis.
| 14
Technical methodology
Production ratio indicators will require some degree of modelling or further analysis meth-
odology to convert observations of abundance to measures of production. Technical meth-
ods are described in the examples used in Annex 2 to 4, and the references cites therein. In
some cases, the use of ecosystem analysis and modelling software such as Ecopath may be
appropriate, but this is not necessarily the case for all ratio measures.
Sensitive to a manageable human activity
Production or biomass ratios at different trophic levels provide a snapshot of the state of
the food web, given the underlying assumptions used in the calculations. It is not possible
to conduct future scenario analyses from such snapshots to evaluate the potential implica-
tions of management measures, but these can be done with some available models, using
system snapshots as initial conditions. Comparative analyses of the same ecosystem in
different time periods, performed using e.g. Ecopath, show that production ratio metrics
are sometimes correlated with changes in human activity.
Relatively tightly linked in time to that activity
It is not clear that responses at the scale of an entire food web could be tightly linked in
time to changes in human activity at all. In fact, the reverse is likely to be true - when
long-lived species are important in the system we can expect considerable lag-times in
response to human activity.
Responsive primarily to a human activity, with low responsiveness to other causes of
change
We cannot expect the status of marine food webs to respond primarily to human activity. It
is clear that they will also respond to environmental changes.
Relevance to Food webs
Ratios of trophic level biomass or production are unambiguously descriptors of the state of
food webs.
Current and historic levels
Ecopath analyses have been carried out for a number of marine regions that will be cov-
ered by the MSFD. However, these have not been performed to a common standard or
based on common criteria for selecting the species or groups to be included, or using uni-
versally accepted parameters. Hence we cannot currently define historic levels that are
valid across regional seas. There is an urgent need for a concerted action to develop the
common standards needed for Ecopath analyses for MSFD regions.
Recommendations for reference levels / limit points
Tentative threshold levels of pelagic: demersal fish biomass have been suggested by some
authors studying food web interactions in specific regional seas, but appropriate levels
cannot be specified at present for all marine regions, or for other ratios.
4.1.2. Criteria 1b) Predator performance reflects long-term viability of components
The abundance of species in the food web will generally be determined by the abundance
of suitable prey taxa on which they can feed. Some species, or groups of species, may play
a significant part in food web dynamics and so their population status will effectively
summarise the main predator-prey processes in the part of the food web that they inhabit.
This metric therefore quantifies the performance of predators through direct population
counts and measurements, which to a large extent are already collected as part of national
| 15
monitoring programmes, and/or planned to support existing and planned programmes (e.g.
OSPAR EcoQOs).
The quantity of food is important as predators that prey upon forage species are sensitive
to fluctuation in prey abundance and can suffer from lack of food resulting from overex-
ploitation or/and environmental changes (e.g. starvation, breeding failure) (Frederiksen et
al. 2007). Food quality is also recognized as critical to the survival of many marine species
including birds (Wanless et al; 2005), mammals (Soto et al. 2006) and fishes (Litzow et al.
2006). For marine birds and marine mammals that are highly dependent on their fish prey
for survival and are keystone predator species in ecosystems (Boyd et al 2006), the re-
quired prey abundance to quantitatively and qualitatively sustain viable populations of
predators should constitute a threshold value. This minimum abundance level of prey nec-
essary to sustain predators can be calculated from existing ecosystem models and could
represent a limit reference point for predator prey interactions within marine ecosystems.
Several studies have shown a connection between seabird survival or breeding success and
the availability (abundance and/or distribution) of key prey species, which mainly are
small pelagic fish (see review by Durant et al., 2004). A particularly relevant example is
the influence of a sandeel fishery in the Firth of Forth, northwestern North Sea, on fledg-
ing success of the black-legged kittiwake, which has been developed into an OSPAR
EcoQO. The breeding success of kittiwakes is calculated using local counts at selected
colonies in Scotland and NE England. The indicator uses the black-legged kittiwake as an
indicator species for the community of predator species that depends on sandeels as an
important food resource. The indicator assumes that if black-legged kittiwakes are unable
to breed successfully for several years in succession, then it is likely that sandeel abun-
dance is low, representing a serious risk of adverse effects on many animal species. The
effect on breeding success is reflected on a yearly basis; the indicator is only triggered
after three years, and benefits of management actions will accrue only in subsequent years.
The breeding productivity at colonies within foraging range of the fishery zone was re-
duced during the period when the fishery was active (Frederiksen et al. 2008), and recov-
ered relative to control areas when the fishery was closed. However, environmental
factors, especially sea temperature were also very influential on fledging success. In terms
of ecosystem management, the results demonstrate that Marine Protected Areas, in this
case a fishery closure, can benefit short-lived pelagic fish stocks and their avian predators.
However, such positive effects require that the regulations of the MPA exclude or restrict
all human activities with negative impacts on the critical resource.
OSPAR has selected the seal population trends indicator for Grey seals (Halichoerus gry-
pus) (declines of less than 10% in pup production) to achieve its ecological quality objec-
tive (EcoQO). Grey seals give birth in terrestrial habitats and are best counted as numbers
of pups produced per year, while harbour seals give birth in intertidal habitats and are best
counted as one-year-old or older seals during the period that they haul-out terrestrially to
moult. This EcoQO would be triggered rather often due to the interannual variations in
numbers of seals (both pups counted or numbers on haul-outs). The probable level of
“alarms” is felt to be too high, and thus a five-year running mean might be applied to these
figures. Such an approach would detect long-term changes in pup production of grey seals
or haul-out numbers of harbour seals. The disadvantage of this is that mortality events,
such as caused by epizootics, would not trigger the EcoQO. ICES felt that this was not a
major disadvantage as large mortality events are already investigated in depth, whereas
more subtle long-term changes might be easily overlooked. The EcoQO as stated in the
Bergen Declaration does not differentiate between sub-units of the North Sea and it is un-
clear whether the EcoQO applies to the whole North Sea population or only to parts of it.
It is not scientifically possible or valid to assess trends for the whole North Sea as there is
a variation in counting methods depending mostly upon the habitat in which the seals are
| 16
giving birth or hauling out. Scientifically consistent trends can be derived for sub-units of
the North Sea, but it should be noted that these sub-units are not necessarily biologically
separate.
There are other potential metrics that could be useful to determine predator performance,
but which are under-developed at this stage. One such metric relates to shifts in the food
web and consequently prey availability, which have been shown to affect body condition
and health of cetaceans and other predator species (e.g. Harwood et al 2000, Bluhm and
Gradinger 2008). For humans and domestic animals it has been shown that a reduction in
nutritional status can lead to reduced reproductive success, affecting age of onset of pu-
berty, fertility, and success in maintaining pregnancies (Gerloff and Morrow 1986) as well
as immune suppression (Landgraf et al. 2005). Reduction in prey availability of marine
mammals and seabirds is also likely to lead to similar adverse effects on health, in particu-
lar causing greater susceptibility to endemic pathogens and increased occurrence of dis-
eases. The health of predators could therefore be used in some circumstances to identify
adverse changes in food webs. Information on the nutritional status of marine mammals
and seabirds can be gained from dead specimens that are collected through stranding net-
works, that have been incidentally by-caught in fishing operations or that can be sampled
live (e.g. seals). It is important to consider if the animals are a representative sample of a
population, as stranded animals alone might have a high proportion of diseased animals
(Murphy et al. 2009). Standard measurements are routinely used to determine body condi-
tion indices of marine mammals and seabirds (Pitcher et al 2000, Read 1990). However,
morphometric indices alone may not be sensitive indicators to changes in condition in
phocid seals and other physiological indices, such as blood variables, have been suggested
(McLaren and Smith 1985, Rea et al 1998).
Another example of a potential index that could be applied to fish is the use of the liver
condition index of Northeast Arctic cod (Gadus morhua) as an indicator of composition of
capelin (Mallotus villosus) and herring (Clupea harengus) in the Barents Sea. Temporal
variation in the liver condition index (LCI) of five length classes of Northeast Arctic cod
was described and compared to the abundance and availability of capelin and herring in
the Barents Sea Yaragina & Marshall (2000). On inter-annual time scales, large and rapid
fluctuations in LCI occurred which were synchronous across length classes. For all length
classes the annual mean LCI was non-linearly related to capelin stock biomass such that
LCI decreased rapidly when capelin stock biomass was below one million tonnes. Liver
condition index and the frequency of occurrence of capelin in cod stomachs were positive-
ly associated. Neither the abundance of juvenile herring in the Barents Sea nor the fre-
quency of occurrence of herring in cod stomachs were positively correlated with LCI.
However, a significant, inverse relationship between the frequency of occurrence of cape-
lin in cod stomachs and total stock biomass of herring was observed suggesting that her-
ring influence cod LCI via predation on capelin. On seasonal time scales, LCI values for
February through July were significantly higher in years of high capelin biomass compared
to years having low capelin biomass. In years of high capelin biomass the proportion of
capelin in the stomach contents of cod showed a peak in March and (or) April.
Both these latter examples (body and liver condition) could be used in the future to devel-
op potential indicators for population status of fish and marine mammals.
4.1.2.1. Recommended indicators of predator performance:
Seal population size and pup production in the North Sea (OSPAR EcoQO (OSPAR,
2005)). Declines in the population size of the harbour seal (Phoca vitulina) or pup recruit-
ment of the grey seal (Halichoerus grypus) indicate poor food supply to seal colonies. The
purpose of the indicator is to maintain healthy populations of seals by triggering manage-
ment actions when needed. Although developed only for the North Sea, the principles can
| 17
be applied to all other European marine waters, and methodological standards are well
documented.
Seabird breeding population size and breeding success in the North Sea (OSPAR EcoQO
(ICES, 2008)). Changes in population sizes are an indicator for important changes in
community structure. Seabird populations may be affected by a range of human activities
although it may take years before these impacts become evident because of the long life-
span and slow reproduction in some seabird species. A change in population trends might
trigger further research to investigate the causes of change, and management might formu-
late "species recovery" or "species action plans". The aim is to maintain a healthy seabird
community. Although developed only for the North Sea, the principles provide valuable
information of food web status and can be applied to all other European marine waters,
and the methodological standards are well documented.
4.1.2.2. Technical evaluation of predator performance indicators;
Easy to understand
Productivity and condition factors of marine animals are easy to understand and to com-
municate.
Data Availability
Data on productivity, condition factors as well as diets of major marine birds and mam-
mals have been collected routinely in some parts of Europe. Thus, where data are avail-
able, the proposed indicators could be easily and accurately measured using data from
monitoring of seabirds, or stranded and by-caught animals as well as data from breeding
colonies.
Technical methodology
Aerial surveys and counts as well as counts from shore are easy to implement and are un-
dertaken routinely by many countries.
Sensitive to a manageable human activity
Marine birds and mammals are typically closely tied to specific geographical locations.
Either because of the location of breeding colonies or their reliance on predictable concen-
trations of prey, they may not be buffered against the effects of longer-term fluctuations in
prey resources. Commercial fishing within the foraging arena of birds and mammal popu-
lations can potentially affect availability of food and have detrimental effects on colonies.
Relatively tightly linked in time to that activity
Seabirds and marine mammals are long-lived species and consequently at the population
level tend to buffer any adverse conditions. Breeding success as well as pup production is
highly sensitive to the local production of food and sometimes induce mass-mortality in
offspring.
Easily and accurately measured
Standard methods in the open sea through line transects have been long developed and
provide accurate estimates, although are not currently available in all marine regions. Seals
can be counted easily while they are on land as well as seabirds on breeding colonies.
Responsive primarily to a human activity, with low responsiveness to other causes of
change
Distribution and abundance of prey, such as pelagic fish, vary substantially with environ-
mental changes and strongly affect survival of seabirds and marine mammals. Fishing ef-
| 18
fects and environmental variability can act in synergy and appear to be difficult to
tangle.
Relevance to Food webs
Top predators are important and emblematic indicators of the overall functioning of the
food-web. They are representative of the general ecosystem state.
Current and historic levels
Historical levels of populations for seals and seabirds are well documented but only few
data are available for marine mammals.
Recommendations for reference levels / limit points
For marine birds and mammals that are highly dependent on their fish prey for survival,
the required abundance to sustain viable predator population of predators should constitute
a threshold value. Minimum viable population sizes are often available for marine birds
and mammals, and represent limit reference points below which populations should not be
driven. Large population increases in seabird and marine mammal populations can also be
detrimental to other components of the food webs and maximum population size can be
defined below which populations should be kept.
4.1.3. Criteria 1c) Trophic relationships that secure the long-term viability of compo-nents
The trophic level (TL) expresses the position of an organism in a food web, and is esti-
mated using diet data. In marine ecosystems, The TL averaged across size/age of a species
population can take any value ranging from 1, for primary producers and other taxa at the
bottom of the food chain, to 5.5, for specialized predators of marine mammals (e.g. the
polar bear) (Pauly et al., 1998). The temporal changes in trophic level of a species or
group of species can indicate progressive changes in prey and can be used to highlight
adverse effects on food web status.
Information about trophic relationships and current prey of species can be obtained
through examination of the diet. Dietary changes can be estimated through isotopic, fatty
acid, stomach content, contaminant analyses and visual observation (e.g. Burek et al.
2008). For marine mammals, this is usually undertaken using stranded or by-caught ani-
mals or in some cases through non-lethal sampling of live animals (e.g. biopsy darting)
(e.g. Krützen et al. 2002). Such an approach can be used to study shifts in prey use of a
species or functional group (e.g. the shift in prey of North Sea harbour porpoise from her-
ring to gadoid species; Santos and Pierce, 2003).
A number of methods which highlight feeding relationships of species in food webs are in
development or are currently applied in some circumstances. There was no consistent
agreement within the Task Group on the extent to which these methods were suitable for
immediate application in EU marine regions. Further evaluation within Regions or Sub-
Regions will be necessary.
In February 2004 the Marine Trophic Index (MTI) was adopted by the Conference of the
Parties to the Convention on Biological Diversity (CBD) as one of eight indicators to mon-
itor achievement by 2010 of a significant reduction in the current rate of biodiversity loss.
The MTI can be calculated from the commercial landings of exploited species (i.e., algae,
invertebrates, fish, marine mammals) (Pauly et al., 1998), as the mean weighted TL of
fisheries landings for a cut-off TL (i.e., TL > value 3.25) (Pauly and Watson, 2005). The
MTI can also be calculated from any measure of biomass or abundance derived from rou-
tine fishery-independent surveys (e.g. data collected from the shelf seas by research ves-
sels: Pinnegar et al., 2002), for different spatial and temporal scales, for example from
| 19
localised ecosystems such as enclosed bays, to larger areas such as the Large Marine Eco-
systems or wider oceanic areas, using annual or seasonal data. Also, the index could be
applied to any assemblage (not just fish) for which there is abundance data for species at
known TL. If the MTI is calculated using fishery landings instead of information from
assemblages, then it will be necessary to interpret the results by investigating changes in
fisheries regulations, technical measures and exploitation strategies. One method to assist
with this task is the Fishing in Balance (FiB) index (for details see Annex 5). Before re-
commending that this indicator is applied operationally throughout European Regional
Seas, further work is necessary to agree generic TL values of fish species (such as those
already provided by FishBase www.fishbase.org) and those for other components (such as
benthic invertebrates) which may also be available.
The dominant prey in diets can be used as a potential index to show temporal shifts in the
main prey consumed. For example, some assessments of diets are already routinely com-
pleted, including stomach content analyses of higher predators (fish, birds and marine
mammals). Additional analyses of diet and associated trophic pathways can be done by
standardizing sample protocols and analysis for isotopic, fatty acid and contaminants for
animals caught, bycaught or stranded. Marine mammal species range from opportunistic to
specialized feeders and the trophic level of their prey also varies. Baleen whales such as
the bowhead or right whale feed on prey such as copepods with a low trophic level. Some
of the toothed whales, such as the Killer Whale, not only feed on squid or fish, but include
higher trophic levels such as other cetaceans or pinnipeds in their diet. Shifts in the food
web and consequently prey availability can have an effect on a number of population pa-
rameters including reproductive success, abundance, distribution, body condition, health,
and mortality. Existing sampling protocols, e.g. within marine mammal stranding net-
works, could be extended for a potential indicator of changes in trophic level of prey, but
this indicator needs more development work before it can be made operational.
4.1.3.1. Technical evaluation of indicators of trophic relationships
Easy to understand
The trophic level of species in a food web describes the level at which a population feeds,
averaged across life-history stages and habitats. Although conceptually relatively simple
the TL varies between individuals in a species and with time, so care must be taken when
applying the concept to time-series data or broadly across eco-regions.
Data Availability
Data describing annual fluctuations in fish population size, either from commercial land-
ings or from fishery-independent surveys, are widely available in European marine waters.
Data quality is dependent on the methods used (such as the gear type and mesh size) and
the accuracy with which landings are recorded. Data for other components (such as marine
mammals, seabirds or benthos) are less frequently available, but can in principle also be
used to track changes in prey (and thereby mean TL of the population).
Technical methodology
The MTI can be easily calculated because it uses a simple measure of abundance (i.e.,
landings, biomass) weighted by the TL value. The quality and reliability of the analysis
and results depend entirely on the accuracy of the TL value. Although such values availa-
ble from online databases such as FishBase, www.fishbase.org, for fishes and SeaLife-
Base, www.sealifebase.org, for other organisms, these make assumptions about the
size/age range of the target populations, and their seasonal feeding ecology. Further work
is necessary to provide reliable TL estimates of species that are applicable in all European
../AppData/Local/Microsoft/Windows/Temporary%20Internet%20Files/OLKBA6/www.fishbase.orghttp://www.fishbase.org/http://www.sealifebase.org/
| 20
seas. The methods used to quantify stomach contents to infer TL can be complex and
costly when relying on isotopic analysis.
Sensitive to a manageable human activity
The TL of a species in a food web can be influenced by adverse human impacts on prey
items, especially for top predators. The MTI, when based on commercial landings, is sen-
sitive to fishing strategies and market values, putting emphasis on the effects of fishing on
the relative abundances of the high-TL organisms (mainly fish), which are generally more
threatened than low TL ones. A strong trend in a long MTI time series is generally affected
by fishing activities whereas year-to-year variability can be the result of both fishing prac-
tices and other causes (e.g. environmental factors, population dynamics).
Relatively tightly linked in time to that activity
The response of the index is on a multi-annual scale.
Easily and accurately measured
It can be easily measured because it uses only a measure of abundance (i.e., landings, bio-
mass) for an array of species and their trophic levels. The estimation of MTI is based on
some assumptions and has drawbacks. Firstly, when using commercial landings data it is
calculated only for the exploited fraction of the ecosystem (i.e., algae, invertebrates, fish,
marine mammals) resulting from fishing strategies and availability and does not take into
account other important biotic components of the food web (i.e., bacteria, viruses, phyto-
plankton, micro-zooplankton, various marine mammals, marine birds and turtles). Thus,
its‟ use assumes that the exploited fraction is representative of wider marine biodiversity.
Secondly, the TL of fish usually changes as fish grow and some species occupy different
trophic levels as they get older. TL can also change from year to year. Thus the use of a
constant TL value might adversely affect the MTI value and the significance and sign of
the trend. Finally, the MTI is sensitive to the TL values used for different species (e.g.
Cury et al. 2005), it might partially reflect changes in the way fishers target different spe-
cies, and does not include discards or illegal landings (which however can be included
should data or estimates become available).
Responsive primarily to a human activity, with low responsiveness to other causes of
change
The strong trend in a long MTI time series based on landings is mainly affected by fishing
activities whereas year-to-year variability can be the result of both fishing practices and
other causes (e.g. environmental factors, population dynamics).
Relevance to Food webs
The feeding relations of marine species, especially those of higher predators, are of direct
relevance to issues related to food web integrity and ecosystem functioning.
Recommendations for reference levels / limit points
The MTI can be linked to a reference point if information is available for periods before
the major industrialization of fisheries. A potential reference point is the mean MTI of
landings (or biomasses) at a time when most stocks were considered to be exploited sus-
tainably.
Despite the availability of some indicators to track rate of change of trophic relationships
in food webs, the development of reference values or reference directions, and acceptable
deviation from these, is complex and needs further work.
| 21
4.2. Attribute 2: Structure of food webs (size and abundance)
One of the simplest means of describing the complex relationships within food webs takes
account of the relative abundance and size distribution of the component species. As food
webs tend largely to be structured by predator prey interactions, the body size of predators,
and the abundance of their prey, will determine the strength and direction of energy flow
through the system. In this section these structural measures are used to identify criteria for
good environmental status of food webs, and suggest simple indicators to record their rate
of change. This attribute links closely with comparable metrics developed to support de-
scriptors related to biodiversity (TG 3) and sea floor integrity (TG 6).